Alex McIsaac here from Northside Ventures!

Each quarter I write a letter to my LPs that includes a section on my market observations. Recently a friend recommended I share my thoughts and observations with the broader community … so here we go!

The first four months of 2025 brought a notable shift in sentiment across markets. After some optimism late last year, public equities pulled back sharply, and consumer sentiment plunged to its second-lowest reading since 1952. President Trump’s sweeping tariffs shocked markets and introduced fresh uncertainty around trade, supply chains, inflation, global capital flows and currency risk. Inflation itself continues to moderate, but high interest rates remain entrenched, and early cracks are appearing in the real economy. Although employment data has stayed resilient, weakening retail sales and growing fears around inflation expectations suggest a slowdown could deepen in the months ahead.

The escalation of US-China trade and technology tensions involving tariffs, GPU export restrictions, and shifting capital controls exacerbated uncertainty across supply chains and innovation markets. As China emerges as a worthy adversary in AI, semiconductors, and defense tech, geopolitics is becoming a more visible force shaping capital allocation and competitive dynamics, and the downstream impact on Canadian tech is something we’ll be watching closely.

In Canada, the election of Mark Carney as Prime Minister last week appears to have introduced a degree of frustration from Conservatives in their party’s inability to sway voters but conversely, a sense of cautious optimism among the business and tech community. Carney brings credibility from his time leading both the Bank of Canada and Bank of England, and his early moves, like scrapping the planned capital gains tax hike, signal a shift toward a more investment-focused, technocratic economic approach. Canadian tech industry voices, many of whom publicly supported the Conservatives, are calling for meaningful action on productivity, procurement reform, and tax incentives that support domestic innovation. There is hope that if the Carney government listens to industry and follows through it could mark a turning point for Canadian competitiveness in tech and venture.

Venture Capital Trends

The early-stage venture market continues to evolve as capital, talent, and ideas concentrate around a narrower set of themes. AI continues to dominate attention, capital is moving earlier, and much of the market is starting to feel homogenous around common problem-solution sets, making original thinking and differentiated access more important than ever.

Here are some of the major venture trends I observed in Q1:

Valuation Bifurcation and Pre-Company Access: Pre-seed round sizes remain relatively stable, typically between $1–3m, but valuations remain highly bifurcated. Repeat founders and standout teams from top accelerators are still clearing $25m to $50m post-money, while the rest of the market clusters around $5m to $12M post-money valuation. Larger funds seem to be moving earlier, driven by compressed risk between Seed and Series A. As a result, top pre-seed firms are sourcing even earlier, spending more time building relationships with founders before a company or idea even exists.

AI Application Layer Saturation: The current environment is marked by growing convergence of capital, themes, and strategy, and the AI application ecosystem, in particular, is becoming increasingly crowded. We’re seeing the nth iteration of the same ideas, especially in vertical SaaS and agentic workflows, and often from strong founders, making it harder to underwrite breakout potential. We believe differentiation will be increasingly tied to data moats, deep domain expertise and a go-to-market distribution edge.

Over-Reliance on Consensus Investing: As early-stage investing becomes more institutionalized, there’s been a drift toward over-optimizing for quantitative signals like milestones, metrics, and resumes that fit a familiar pattern and consensus. While metric-driven decisions and milestone investing have their place, they cannot fully capture the nuance of early-stage potential that is rooted more deeply in the founding team. We’re seeing fewer investors lean into unconventional founders or novel markets unless they slot neatly into pre-existing frameworks, leaving real opportunity on the table. At Northside, we’re increasingly drawn to high velocity outlier founders who defy consensus norms — for example we recently backed two companies (still in stealth) each founded by University of Waterloo dropouts, each aged 19 and 23. Both founders are special individuals with incredible build velocity, an overdrive mentality and massive ambitions, which got us excited to invest.

Liquidity Reset and the Rise of Secondaries: Trump tariffs and market volatility have resulted in fading hopes of a broad IPO or M&A rebound and it’s becoming clear we’re in a new structural reality. IPOs are likely to be fewer and later, reserved for mature, capital-efficient companies and in this environment, secondaries are no longer a side door — they’re becoming the primary liquidity path for early investors. Founders and early investors in growth stage companies are already securing meaningful liquidity through structured tenders and private sales. For smaller funds, this shift raises the importance of building relationships with later-stage buyers and actively managing liquidity within private markets. Northside is still a young fund in its 3rd year of the investment period, but liquidity is always top of mind.

Category Trends We’re Watching

This quarter reinforced that the AI stack across models, infrastructure, and applications is evolving faster than ever. The gap between research and deployment is shrinking, and a new set of technical and geopolitical dynamics is reshaping where value accrues.

Continuing on my post from last month titled “Tech Trends I Observed in Q1 2025”, here are some of the most notable category trends we’re watching today:

Agentic AI Goes Commercial: The most important shift recently has been the full emergence of agentic AI systems. 2025 is proving to be the year when AI moves beyond chat interfaces to autonomous action — systems that fetch live data, call APIs, and orchestrate tasks across real-world applications without human prompts. Engineering teams are leaning heavily into AI coding assistants such as Cursor, which surpassed $200m ARR this March, or Windsurf, which OpenAI acquired this week for $3 billion.

The AI Infrastructure Rebuild Is Underway: Legacy systems built for static, human-triggered workflows are proving inadequate for AI-native operations. Founders are now rewriting infrastructure stacks from the ground up for continuous learning, autonomous decision-making, and dynamic data flow — paving the way for a new generation of AI-first software. This is an area we’re especially excited about because of the huge market potential as SMBs and enterprises adopt AI en masse.

Generative Worlds on the Horizon: We see the early signs of generative worlds becoming the next frontier, following the generative image boom of 2023 and the generative video breakthroughs of 2024. Stability AI’s advancements in 3D scene and camera angle generation from 2D images are opening fresh possibilities in gaming, virtual production, and digital content creation. One notable example is Moonvalley (fka Contentfly), a company I invested in pre-seed and pre-YC while at my former fund, Global Founders Capital, who released their AI-video generating model, Marey, trained exclusively on owned or fully licensed source data, and now has over $113m in backing from General Catalyst, Khosla Ventures and Bessemer Venture Partners.

China’s Acceleration Across the Stack: China’s presence in AI became harder to ignore this quarter. Labs like DeepSeek and Alibaba released near-frontier language and video models, while Huawei made meaningful progress in AI silicon. Despite export controls, Chinese firms are advancing quickly across the model, chip, and tooling layers. With a national push to replicate NVIDIA’s computing platform CUDA and close remaining software gaps, the global AI race is no longer US-dominated. The long-term implications for compute infrastructure and software competition will be hard to ignore.

Deeptech Hardware Gains Ground: Deeptech hardware is emerging as a more defensible venture category. As software development becomes increasingly democratized through tools like Replit and Cursor, which are turning natural language into the new language of programming, companies building deeply technical hardware spanning robotics, semiconductors, and industrial systems will be more difficult to replicate quickly, and are opportunities for startups to build deeper long-term moats.

Closing Thoughts

The first four months of 2025 reaffirmed that venture is becoming more uniform and more crowded around common themes, playbooks, and founder profiles. But markets don’t reward consensus. The most enduring companies are built by outliers — those with a differentiated view of where the world is going and the drive to build toward it early.

At Northside, our focus is unchanged: backing visionary founders when nothing else exists. We aim to be their earliest partner, collaborating with conviction, listening deeply, and helping them shape the future through clarity, trust, and thoughtful support.